چکیده:
Integrating flexible job-shop scheduling problem (FJSP) with
preventive maintenance (PM) is pondered in this paper. Minimizing the
makespan for scheduling part and minimizing the system unavailability
for maintenance part are simultaneously under consideration. For doing
it, the assignment of n jobs on m machines in production side and
executing the PM actions at appropriate time intervals in maintenance
part are carried out at the same time. Also, for carrying out the
maintenance side, reliability model is employed. Moreover, number of
maintenance actions and maintenance intervals are not fixed in
advanced. In order to ensure of obtained results, two multi-objective
evolutionary algorithms (NSGA-II and NRGA) are compared. Besides,
these genetic algorithms were hybridized with both well-known
composite dispatching rule (CDR) and active scheduling and then
compared as a two new evolutionary algorithms named CDRNSGA-II
and CDRNRGA. Furthermore, the algorithms are compared with more
than 4800 test instances.
خلاصه ماشینی:
"1-29 Scheduling integrated flexible job-shop problem with preventive maintenance to minimize makespan and system unavailability E.
Fatemi Ghomi1,* Abstract Integrating flexible job-shop scheduling problem (FJSP) with preventive maintenance (PM) is pondered in this paper.
Key words: Scheduling, Flexible job shop, Preventive maintenance, Genetic algorithm, Reliability, Multi-objective optimization.
Subjected to the Figure 2 where the integrated production and maintenance activities are portrayed and considering the chosen policy and rational strategy, the unavailability of machines and system are shown in Table 2.
Chromosome evaluation: Step 1) Assign the jobs to the machines according to the FJSP Step2) Change the scheduling of FJSP to active FJSP (for algorithms accompanied by active scheduling) Step 3) Calculate the makspan Cmax Step 4) Insert the PM tasks in production scheduling using the fixed interval policy and rational strategy Step 5) Reevaluate the makespan after inserting the PM activities Step 6) Calculate the system unavailability.
The CDR algorithm Figure 11 shows the algorithm designed to decode an OOMS to a feasible and active schedule for a Flexible Job-Shop Problem which is introduced in Ho et al (2007).
Computational results To compare four algorithms used to solve our problem, we reviewed standard flexible job-shop test problems from Barnes and Chambers (1996), Brandimarte (1993), Dauzére-Pérés and Paulli (1997), and Hurink et al.
Conclusions and future works A bi-objective integrated model to solve the joint production and maintenance scheduling problem in the FJSP case is pondered in this paper.
Bi- objective optimization algorithms for joint production and maintenance scheduling: application to the parallel machine problem."